KR20010082984A - A system for retrieving world wide web and a method for storing, viewing and using the search result - Google Patents

Abstract

PURPOSE: A system for searching a world wide web page and a method for storing, viewing, and using the search result are provided for a user to access to wanted information accurately and rapidly using hyperlink information on a world-wide-web and a tree-structure of a web page. CONSTITUTION: A hyperlink index word extracting unit(1500) extracts hyper links indicating a specific web page from a distribution database(1100), and selects index words from a link text of hyper links and forms an index gathering(default URL) responding to the corresponding web page and stores the index gathering in a hyperlink file(1600). An index word expanding unit(1400) reads one index gathering from the hyperlink file(1600) and calls a web page having the index gathering as an address from the distribution database(1100) and selects an important index word from the corresponding web page and appends the important index word to the index gathering, thus the index gathering is expanded. An index word weight value embodying unit(1300) receives the index gathering from the index word expanding unit(1400) and extracts a web page having the index gathering as an address and position information of sub pages being connected to a depth 'n' of a tree structure, and calls the corresponding sub page from the distribution database(1100) based on the position information. A query word processing unit receives a user query word(1220) through a user interface(1200) and extracts an index word by analyzing the query word and reads position information gathering responded to the index word extracted from a conversion file(1700) and processes the information gathering.

Description

A SYSTEM FOR RETRIEVING WORLD WIDE WEB AND A METHOD FOR STORING, VIEWING AND USING THE SEARCH RESULT}

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a search engine that facilitates web surfing on the Internet, and in particular, by using hyperlink information on a world-wide-web page and a tree-structure of a web page. The present invention relates to a retrieval system that enables a user to accurately and quickly access desired information.

Conventionally, techniques for parsing, indexing, and searching documents in a database are well known.

In recent years, distributed databases have emerged in the form of a world-wide-web. The database document on the web is in the form of pages accessible via the Internet. Thus, anyone who has a communication link to the Internet can access a web page.

Web pages are distributed on millions of different computer systems around the world. Internet users wish to be placed on a particular web page that contains information of interest.

The method of accessing these web pages via the Internet is as follows.

When the user specifies the location information (URL: uniform resource locator) of the web page containing the information to be viewed by the WWW browser, the document indicated by the location information is transmitted from the website to the user's computer. This is converted into a predetermined format by a web browser of the user's computer and displayed on the screen.

However, in this way, if the user does not know the URL, the target document cannot be accessed. In addition, the amount of information on the Internet is enormous, and simply following the link relationship between documents does not necessarily mean that the target document can be reached. Therefore, a search engine that provides a document retrieval process is required.

When the search engine receives a search request from a user computer, the search engine starts a search program through a common gateway interface (CGI). The search program inquires into the index database and converts the search results received from the database into HTML (Hype Text Markup Language) format. Then, the generated HTML document is sent to the user computer. In this way, the user can browse various URL information.

However, such a conventional search engine, for example, ALTAVISTA (a representative search engine that receives a keyword by searching for it) and the like extracts all index words from the corresponding web page documents on a large number of web pages existing on the Internet. Index web pages.

Therefore, index words (hereafter defined as stop words) that do not represent the characteristic information of the web page are also extracted, which increases the size of the index database, takes a long time to build, and makes it difficult for users to access the information exactly as they want. .

The present invention relates to an indexing system that adds a conventional content-based search based on structure-based search in indexing a web page.

It is a first object of the present invention to provide a method of selecting and indexing only a web page of a URL structure that can access a web page even if a file name is omitted in a hierarchy of URLs.

A second object of the present invention is to extract an index word from the link-text of this hyperlink when at least one or more hyperlinks in the plurality of external web pages point to the same specific page and extract the index word. It provides a technique for indexing the web page.

It is a third object of the present invention to provide a method of selecting an index word that defines a web page in consideration of a tag structure on a web page, word affinity, and registration of a general vocabulary dictionary.

A fourth object of the present invention is to provide a method for assigning index word weights by checking the frequency of index word occurrences not only in the web page but also in subpages linked to the web page in indexing a specific web page with specific index words. It is.

It is a fifth object of the present invention to provide a method for non-stop displaying a web document weighted with the highest weight for a specific query on a user computer.

A sixth object of the present invention is to provide a method of storing the address of a web document retrieved from a search engine server in a specific server on the Internet linked with a user computer and easily utilizing the same.

An indexing system of a worldwide web page as a first aspect of the present invention for achieving the above object comprises a communication interface for acquiring a corresponding web page from the location information (URL) of the worldwide web page via the Internet, and this communication interface. Hyperlink retrieval means for sequentially extracting all the hyperlinks on the webpage obtained by step by step, and the webpage can be accessed even if the file name is omitted or omitted in the uniform resource locator (URL) structure. Hyperlink selecting means for selecting only hyperlinks having a default URL, and collecting means for collecting hyperlinks having the same default URL from the World Wide Web database, collected by the collecting means. Index word extracting means for extracting an index word from the link text of the hyperlinks.

The index word extracting means includes: an index word extracting module for forming a plurality of " location information: index list " from hyperlinks having the default URL, a first storage file for storing and storing the " location information: index list " By searching all index indexes corresponding to the specific location information stored in the first storage file, counting the frequency of occurrence of each index word, and weighting the index word from the frequency of occurrence, " location information: index index (weighted) list. An index module for forming ", a second storage file for storing the " location information: index list ", and " location information: index list " list of the second storage file " And a third storage file for storing the " index: location information (weighted) list ".

Therefore, the web page indicated by the default URL can be indexed by the extracted index words.

In general, the hypertext of a worldwide web page contains a hyperlink that links the web page and the web page so that users can browse and search the web page in a non-sequential way by following the link. The hyperlink includes text for representing the web page with location information (URL) indicating the web page. Therefore, when the index word is extracted from the link text of the hyperlink, it is possible to extract the index word more accurately with less index work than when extracting the index word on the text of the web page.

A worldwide web page indexing system as a second aspect of the present invention obtains a web page from a distributed database on the Internet based on location information (URL), and extracts an index word from the link text of a hyperlink existing on the web page. Hyperlink indexing means for forming " location information: index set ", a hyperlink index file for storing the " location information: index set ", and a " location stored in the hyperlink index index file. Information index indexing means for reading the "information index index", extracting a key index word from the text on the web page corresponding to the read position information, and expanding the index word set, and to a depth N connected to the web page and the tree structure. Acquire a subpage of, search the obtained subpage and the webpage and index the post-expansion index An index word weighting means for weighting index words of the index set according to the frequency of occurrence by checking the frequency of occurrence of each index word, and " location information: index word (weighted value) weighted by the index word weighting means. Contains an index file to store the "set".

A second aspect of the present invention provides a function of extracting a key index word from a document on a web page to be indexed, adding the extracted index word to an index word extracted from hyperlink information, and extending the index word, and weighting the extended index word. The web page to be indexed and the subpage of the web page are used to provide more precise and accurate search.

The index word weighting method according to the third aspect of the present invention provides a method for indexing a web page using a distributed database on the Internet, such that a plurality of index words that characterize a particular web page and location information of the web page correspond to each other. An index word set ", and stores the same, and retrieves the web page indicated by the position information of the" location information: index word set "and position information about subpages up to a depth N connected in a tree structure. Obtains a subpage indicated by the location information extracted from the distributed database, counts the frequency of occurrence of the index words of the " location information: index set " This is achieved by assigning a weight to the frequency of occurrence.

According to a fourth aspect of the present invention, a method of viewing a search result analyzes a query input through a user computer to extract at least one index word, and matches the index words from the index database using the extracted index words. Searching the plurality of web page location information, sorting the plurality of searched location information in descending order according to the weight, and delivering the highest ranking location information among the plurality of location information arranged in descending order to the web browser of the user's computer. This is accomplished by having a web page compliant with the location information displayed on the user's computer nonstop.

Further, the fourth aspect of the present invention may include displaying on the user computer a list of the plurality of location information arranged in descending order.

In this case, it is preferable that the web page corresponding to the highest position information and the list of the position information are simultaneously displayed on one screen.

A method of storing a search result as a fifth aspect of the present invention comprises the steps of: a search engine server retrieving document location information (URL) lists associated with a query word transmitted from a user computer from an index database; Reordering by a processing method, transmitting the searched document location information list to the user computer by the search engine server, displaying the location information list on the screen of the user computer, and displaying the location displayed on the user computer. Selecting at least one or more desired location information from an information list, transmitting the selected at least one location information to a specific server linked to a user computer, and storing the location information transmitted by the specific server of the user Including saving to the area .

In addition, a method of storing a search result as a sixth aspect of the present invention includes the steps of: transmitting, by a search engine server, location information of a search document to a user computer; and a web browser of the user computer recalls the search document based on the location information. And displaying the location information of the displayed search document to a specific server linked with the user computer according to the user's decision, and transmitting the location information transmitted by the specific server to the corresponding user. And storing in a storage area of the.

As a method of using a stored search result as a seventh aspect of the present invention, a plurality of user computers and a server computer are connected by using the Internet, and the server is provided with a plurality of storage areas for the plurality of user computers linked with the server. In a storage environment, a plurality of web document location information lists transmitted from a user computer are stored for each user in the storage area, wherein the user accesses the server computer using his computer, and stores the server computer. Retrieving a plurality of location information lists from a storage area for the user in the area and displaying them on the user's computer, and selecting desired location information from the displayed location information list to display a web page corresponding to the location information. Accessing.

Therefore, the user can easily surf the web simply by connecting to the server computer.

Other objects and advantages of the invention will be described below and will be appreciated by the practice of the invention. Furthermore, the objects and advantages of the present invention can be realized by means and combinations indicated in the appended claims.

The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate the presently preferred embodiments of the invention and, together with the description of the following preferred embodiments, serve to explain the principles of the invention.

1 is an overall system configuration for explaining the search system of the present invention.

FIG. 2 is a diagram for explaining a state where a specific web page and a plurality of external web pages are connected by a hyperlink.

3 is an internal block diagram of the search engine of the first embodiment according to the present invention.

4A to 4D are flowcharts illustrating a process of building an index database using the search engine of FIG. 3.

FIG. 5 is a flowchart for describing a process of processing a query input from a user computer by the search engine of FIG. 3.

6 illustrates a screen configuration of a web browser in which search information provided from a search engine of the present invention is displayed.

7 is an overall block diagram of the search engine of the second embodiment according to the present invention.

8A to 8B are detailed internal configuration diagrams of the search engine of FIG. 7.

9 illustrates a state in which a plurality of web pages are connected in a hierarchical tree structure on the web.

FIG. 10 is a diagram for describing a process of assigning a weight for each index word from a web page to be indexed by the search engine of FIG. 7 and subpages of the web page.

11A through 11D are diagrams for describing a process of building an index database using the search engine of FIG. 7.

12 is a system configuration diagram for executing a method of storing and using a search result according to the present invention.

<Description of the symbols for the main parts of the drawings>

1000: search engine 1100: distributed database

1200: user interface 1300: indexed weighting

1400 index index expansion unit 1500 hyperlink index extraction unit

1600: hyperlink file 1700: conversion file

1800: index file 1900: query processing unit

Hereinafter, with reference to the accompanying drawings will be described embodiments of the present invention;

1 illustrates a distributed computer system 100 that includes an index database.

Distributed system 100 includes a client computer 110 that is connected to a server computer (website) 120 via a network 130. This network 130 may use an Internet communication protocol (IP) that allows a client to communicate with a server.

The client computer 110 may be a personal computer, a workstation or a computer system larger or smaller. Each client typically includes one or more processors, memory, and input / output devices. The server also has a similar configuration to the client.

In most cases, however, serversite 120 includes a number of computers connected by a local network. In effect, the network includes hundreds to thousands of individual computer networks.

Although the client computer is shown separately from the server computer, a single computer can serve as both a client and a server at the same time.

In operation of this distributed system the user of the client computer wishes to access the information record 122 stored by, for example, the World Wide Web server. The information record 122 may be in the form of a web page 200. Web page 200 may be a data record such as simple content text information or more complex digitally encrypted multimedia content such as software programs, graphics, audio signals, video, and the like.

The client includes a web browser program 112, such as a navigator or explorer, to locate the web page. This browser program 112 allows the user to enter the address (or location information) of the particular web page to search. The address (or location information) of the page is specified as a uniform resource locator (URL). In addition, the browser program makes it possible to access another page by clicking on a hyperlink of the obtained web page when the page is obtained. These hyperlinks automate entering the URL of another page after obtaining the page.

A search engine 140 is provided to identify pages of interest to the user from the millions of available web pages on the web.

The search engine 140 is a multiprocessor (P) 142, a memory (M) 144, an index database 146 and a network interface 148 and a driving program connected to each other by a high-speed communication bus 143. Is done. Of course, the present invention is not limited to a search engine composed of multiple processors connected to each other by a high speed communication bus.

Hereinafter, the configuration of the preferred embodiments of the search engine according to the present invention with reference to the accompanying drawings.

Example 1

2 is a view for explaining the technical concept of the search engine according to the present embodiment.

The present embodiment is characterized in that indexing of a specific web page is performed from link text of at least one hyperlink pointing to the specific web page, rather than mainly from the text document of the web page.

In more detail with reference to FIG. 2, the web page 10 to be indexed is composed of a location information 11 and a text document 12. In addition, the web page 10 is connected to a plurality of external web pages 20 and hyperlinks 21.

This hyperlink 21 largely includes positional information 21a (URL) and link text 21b. The location information 21a is defined as the URL 0 of the web page 10 to be accessed via this link. The link text 21b includes a representative phrase for summarizing and representing the text document 12 of the web page 10 indicated by the location information 21a.

Therefore, when extracting index words from the link text 21b (link texts 1, 2 ... n) of the hyperlinks 21, these index words are displayed on the web page indicated by the hyperlink position information 21a. 10) This will be a word representing the information on the image.

In the conventional search engine, all possible index words are extracted from the text 12 of the web page 10 to be indexed and the corresponding web page is indexed. However, although the text 12 of the web page 10 includes words representing the information of the web page, most of them are composed of words (hereinafter, defined in a stop word) that do not represent the web page.

Therefore, when the web page is indexed by the same method as before, the indexing operation is enormous and the accuracy of the search is inferior.

That is, the present embodiment extracts index words from the link text 21b of one or more hyperlinks 21 indicating the web page 10 to be indexed as shown in FIG. Can be.

Hereinafter, the concept of FIG. 2 will be described in detail with reference to FIGS. 3 to 5.

3 shows the internal structure of the search engine of the first embodiment.

<Building Index Database>

As shown in FIG. 3, the search engine of the present embodiment includes a URL input and storage unit 40, 41, 42, 43, 44, a communication interface 30, a hyperlink extraction module 50, and a hyperlink analysis module ( 60, index word extraction module 70, hyperlink file 72, index file forming unit (73, 75), conversion file forming unit (76, 78), query processing and ranking module (80) and query module ( 90).

The URL input and storage unit comprises a user interface 40, a first URL storage file 42, a storage file reset and URL assignment module 43, and a second URL storage file 44.

The user interface 40 is a means for the administrator of the search engine server to input an initial URL to the search engine server for building an index database. Here, the initial URL is preferably the root URL 41. Here, the root URL refers to a URL in which a ancestor page corresponding to a hierarchical upper level of the corresponding URL does not exist in the hierarchical tree structure on the web. The tree structure also includes a graph structure between webpages that reside in the directory containing the webpage indicated by the root URL and subdirectories that can be reached along hyperlinks in the webpage indicated by the root URL.

At this time, the administrator of the search engine server inputs at least one or more root URLs 41 through the user interface 40.

The first URL storage file 42 is a storage means for storing the initial root URL 41 input from the user interface 40 and the root URL 62 transmitted from the hyperlink analysis module 60 described below. to be.

The storage file reset and URL assignment module 43 resets the second URL storage file 44, reads one root URL from the first URL storage file 42, and then stores the second URL storage file 44. Program module that executes commands to be stored in.

The second URL storage file 44 stores the root URL assigned by the storage file reset and URL assignment module 43 and stores the subpage URL 61 transmitted from the hyperlink analysis module 60 to be described later. Storage means for storing (the second URL storage file may exist in RAM).

The communication interface is a means for interfacing the search engine 140 and an external website 120 (web server) through a network 130 such as the Internet.

3, this communication interface is defined as a web page acquisition module. That is, the web page obtaining module 30 reads one URL 45 from the second URL storing file 44, and the web page 200 indicated by the read URL 45 is connected to a network such as the Internet. The program module obtained from the web server 120 via the 130.

The web page acquisition module 30 transmits the web page 200 acquired through the Internet 130 to the hyperlink extraction module 50 described later.

The hyperlink extraction module 50 is a program module that sequentially extracts all the hyperlinks existing on the web page 200 and sequentially delivers them to the hyperlink analysis module 60 which will be described later. Here, the term "extract and transfer the hyperlink" means determining whether a hyperlink exists on the web page 200, and obtaining a web page when the hyperlink does not exist on the web page 200. If a hyperlink exists on the web page 200, the module 30 is notified, and the hyperlink which extracts the position information and the hyperlink text 51 of the hyperlink later is extracted one by one so as not to overlap. It indicates to each of the analysis module 60.

The hyperlink analysis module 60 is a program module for analyzing the location information of the hyperlink transmitted from the hyperlink extraction module 50. That is, by analyzing the URL of the hyperlink delivered from the hyperlink extraction module 50, it is determined whether the URL includes the root URL portion read from the first URL storage file, and the URL is the default ( default) URL is also determined.

Here, the root URL is used in the same meaning as the above definition, and the subpage URL is defined as a URL indicating a lower page that is hierarchically dependent on the root URL in the hierarchical structure of the URL.

In addition, the default URL refers to a URL configured to access a web page indicated by the URL even if the file name is omitted or omitted from the structure of the URL.

In general, a URL has a hierarchical structure such as "protocol: // computer address / directory name / file name". Here, the protocol refers to a communication protocol used to transmit a resource represented by a URL such as http (hypertext transfer protocol) and ftp (file transfer protocol) in a communication network, and the computer address refers to the address of the other computer where the resource is stored. The directory name is the directory on that computer.

For example, in "http://www.microsoft.com/Windows95/help.html", "www.microsoft.com" represents a computer address, "Windows95" represents a directory name of the computer, and "help. html "represents the file name where the resource is located in that directory.

Therefore, in order for a web browser to access a specific web page, in principle, the full address (up to the file name) of the web page must be entered. However, if the URL of the web page is the default URL, the web browser can access the web page even if the file name is omitted among the entire addresses of the web page.

The hyperlink analysis module 60 stores the root URL 62 in the first URL storage file 42 and the subpage URL 61 through the routines of steps S9 to S24 of FIGS. 3A to 3B described later. 2 URL storage file (44).

In addition, the hyperlink analysis module 60 determines whether the URL indicated by the extracted hyperlink is the default URL, and selects only the hyperlink having the default URL to describe the corresponding "location information and link text" 63. Transfer to the index word extraction module 70.

The index word extracting module 70 extracts an index word from the link text of the hyperlink having a default URL, and uses the extracted index words to &quot; location information: index word list &quot; 71 (URL: Index = a, b, c). Is a program module that forms

The " positional information: index list " 71 (URL: Index = a, b, c) formed by the index word extracting module 70 is stored in the hyperlink storage file 72. As shown in FIG. Here, "accumulated" means that a plurality of index index lists corresponding to the same positional information can be formed and stored. That is, for example, a plurality of index words such as (a, b, c), (d, e, f), (a, e, g), (a, e, g), etc. for a specific default URL (0). Lists are matched, and accordingly, "URL (0): (a, b, c)", "URL (0): (d, e, f)", and "URL (0): (a, e, g)" A plurality of &quot; location information: index word lists &quot; such as &quot; URL (0): (a, e, g) &quot; can be stored in the hyperlink storage file.

In this manner, when the extraction of the hyperlink and the extraction of the index word from all the root URLs and the subpage URLs of the first URL storage file 42 are completed, the hyperlink storage file 72 includes millions of "location information: Index index list ”is stored.

The index file forming module 73 is a program module for forming the index file 75 to be described later by assigning a weight to each index word from a plurality of "location information: index word lists" having the same position information. That is, the index file forming module 73 collects the &quot; location information: index list &quot; having the same location information from the hyperlink storage file 72, and checks the occurrence frequency of each index word from the plurality of index index lists thus collected. The weight is given according to the frequency of occurrence. For example, if the index word a appears 100 times in the index list corresponding to the default URL 0, the weight of the index word a is 100. By repeating this operation, each index word is weighted, and one index word (weight) list corresponding to one URL [URL (0): a (w1), b (w2), c (w3) ) Can be formed. The thus-formed "location information: index word list" 74 (URL (0): a (w1), b (w2), c (w3) ...) is stored in the index file 75 described later. .

When all &quot; positional information: index word (weighted) list ”are stored in the index file 75, the converted file is formed by the conversion file forming module 76 described later.

The conversion file forming module 76 converts the "location information: index word (weighted) list" stored in the index file into the "index word: location information (weighted) list" 77 (a: URL (w1), URL (w2), URL). (w3) Program module 78 for converting to ....

The "index word: position information (weighted) list" 77 (a: URL w1, URL w2, URL w3) is stored in the conversion file 78, thereby creating an index database.

An index database construction process based on the above-described configuration will be described in detail with reference to the flowchart of FIG. 3.

When the search engine server administrator of the present invention inputs at least one web page root URL 41 through the user interface 40 of the server computer 140, the root URL 41 is the first URL storage file 42. Are stored in. (Step S1)

The storage file reset and URL assignment module 43 reads one root URL from the first URL storage file 42. (Step S2) At this time, if it is impossible to read the root URL from the first URL storage file, the process proceeds to step S30, and if the reading is possible, the process goes to step S4. (Step S3) Here, the inability to read the root URL from the first URL storage file means that there is no information of the root URL to be processed anymore.

If the root URL can be read in step S3, the storage file reset and URL assignment module 43 resets the second URL storage file 44, and reads the read root URL to the second URL storage file 44. Save it. (Step S4)

The web page acquisition module 30 reads one URL 45 from the second URL storage file 44. (Step S5) At this time, if the URL 45 cannot be read from the second URL storing file 44, the flow returns to step S2, and if the reading is possible, the process goes to step S8. (Steps S6, S7) Here, the inability to read the URL from the second URL storage file means that the extraction of the hyperlink information and the index word extraction for all the URLs present in the second URL storage file have been completed. This means that a new root URL is allocated from the first URL storage file and a reset of the existing URL is required.

If the URL can be read in step S6, the web page obtaining module 30 obtains the web page 200 indicated by the URL from the web server 120 on the Internet 130 (step S8).

When the acquisition of the web page is completed, the web page is delivered to the hyperlink extracting module 50, and the hyperlink extracting module 50 sequentially extracts the hyperlink from the web page (step S9).

At this time, if there is no hyperlink to be extracted on the web page, the process returns to step S5, and if there is a hyperlink, the process proceeds to step S11. (Step S10) Here, if there is no hyperlink to be extracted on the web page, if there is no hyperlink to be extracted from the web page at all and the extraction of all hyperlinks existing on the web page is completed. Includes cases.

If it is possible to extract the hyperlink from the web page in step S10, the hyperlink extraction module 50 obtains the link URL and the link text from the extracted hyperlink information. (Step S11)

The root URL is extracted from the link URLs obtained in step S11, and it is determined whether the root URL matches the root URL in step S2. (Step S14)

At this time, if the root URLs match with each other (step S15), it is determined whether or not the link URL obtained in step S13 exists in the second storage file. (Step S16) At this time, if the link URL exists in the second storage file, the process proceeds to step S24 to determine whether the link URL is the default URL, and if not, the process proceeds to step S20 and transfers the link URL to the second storage file. Save and go to step S24.

If the root URLs do not coincide with each other in step S15, it is determined whether the root URL obtained in step S13 exists in the first storage file. (Step S17)

In this case, if the root URL exists, the process proceeds to step S24 to determine whether the obtained link URL is the default URL. If not, the process proceeds to step S23, and the root URL obtained in the step S13 is stored in the first storage file. Proceed to S24.

Steps S13 to 23 are routines for explaining a process of extending URL information of the first and second URL storage files by extracting another root URL and a subpage link URL from one root URL. At this time, it is determined whether the extracted root URL and the subpage link URL exist in the first and second storage files in order to prevent storage of duplicate URLs.

In step S24, the hyperlink analysis module determines whether the link URL obtained in step S11 is a default URL. At this time, if the link URL is not the default URL, the process returns to step S9 (step S26), and another link URL is extracted from the web page, and if it is the default URL, the process goes to step S27.

Here, the reason for selecting only the default URL among the extracted link URLs is that the web page indicated by the default URL tends to represent the sub-web pages and thus contains more important information than the sub-web pages. This is because it is possible to prevent duplication of similar search results and to move freely from the web page indicated by the default URL to the sub-web page.

In step S27, the index word extracting module extracts all possible index words from the link text of the corresponding default URL to generate "location information: index list" [URL (0) :( a, b, c)], and the hyperlink storage file. Is accumulated (step S28), and the flow returns to step S9.

If the root URL to be read from the first URL storage file no longer exists in step S2 in the above-described routines of steps S2 to S29, the process proceeds to step S30 to execute the index file and the conversion file generation routine. do.

In step S30, the index file forming module reads a plurality of &quot; location information: index list &quot; having the same URL from the hyperlink storage file, checks the occurrence frequency of the index words therefrom, and weights according to the occurrence frequency for each index word. To generate a "positional information: index list" (URL (0): a (w1), b (w2), c (w3) ...).

A plurality of &quot; positional information: index list (weighted values) &quot; (URL (0): a (w1), b (w2), c (w3) ...) generated in this way is stored in the index file. (Step S31)

In step S32, the conversion file generation module reads &quot; location information: index list (weighted) list &quot; [URL (0): a (w1), b (w2), c (w3) ...)] from the index file. An index database is formed by converting the "index word: position information (weighted value) list" [a: URL1 (w1), URL2 (w2), URL (w3) ...] into a conversion file. (Step S23)

<Search the index database>

The search for the index database (or the conversion file) is performed similarly to the conventional method, which will be described below with reference to FIGS. 3 and 5.

A user who wants to access a specific web page enters a query into the search engine 140 of the present invention. (Step S41)

This query can be a keyword or natural language linked by one keyword or multiple operators.

When the query 91 is input, the query module 90 interprets the query 91 and extracts a related index to form the query request 92. This query request 92 is forwarded to the query processing and ranking module 80.

The query processing and ranking module 80 searches for "index: location information (weighted) list" having the same index as the query requested index in the conversion file (index database) 78. (Step S42)

The searched "index word: location information (weighted) list" [index word (a): URL1 (w1), URL2 (w2), URL3 (w3) ...]] is transmitted to the query processing and ranking module 80, and the query The processing and ranking module 80 processes the retrieved location information according to the index word coefficient and operator definition in the query word, and then sorts the final location information search result in descending order according to the high and low of the weight, and then the query module 90 (Step S43).

For example, the "location information list" transmitted to the query module is as shown in (1) below.

"URL2 (w2), URL1 (w1), URL3 (w3) ..." (1)

(Where w2> w1> w3 ...)

The query module, which has received the "location information list" arranged in descending order, causes the corresponding search result to be displayed on the web browser screen of the user's computer as described later based on the location information list. (Steps S44, S45)

<Display of search results>

Hereinafter, a viewing method of a search result according to the present invention will be described with reference to FIGS. 5 and 6.

When the location information list 81 arranged in descending order is transmitted to the query module 90, the query module 90 divides the web browser screen into a plurality of frames so that different documents or files can be displayed in each frame. do.

That is, the query module 90 enables the web browser to retrieve a web page corresponding to the highest weighted position information (URL) from the web server on the Internet and display the web page on the screen of the web browser on a non-stop basis. The first frame information, the second frame information for displaying the location information list 81 on the web browser screen of the user computer, and the third frame information for displaying the search box or the advertisement display unit on the web browser screen. Pass it to the browser.

When a document file containing a plurality of frame information is transmitted from the query engine module 90 of the search engine to the web browser of the user's computer, the web browser interprets the corresponding document file and constructs a web browser screen using the frame information.

6 illustrates a state in which a web browser configures a corresponding browser screen based on frame information transmitted from a search engine.

The web browser screen 150 is composed of largely divided frames, wherein one frame 154 displays the first frame information, and the right side of the frame 154 displays the second frame information. One frame 153 is positioned, and the other frame 155 for displaying the third frame information is positioned at the top of the frame 154.

Of course, the browser screen of the present invention is not limited to being composed of frames of detail, but may be configured with more frames or fewer frames.

In addition, the mutual positional relationship between the frame 154 and the remaining frames 153 and 155 is not necessarily limited to the configuration as shown in FIG. That is, the position of each frame may be interchanged. For example, the frame 155 may be located at the bottom of the frame 154, or the frame 153 may be located at the left side of the frame 154.

The web browser retrieves a web page corresponding to the position information from a web server on the Internet based on the highest ranking position information transmitted from the search engine and displays the web page non-stop in a frame 154 of the browser screen. S45 of FIG. 5 Therefore, since the user can be provided non-stop a web page most closely associated with the desired information, it is possible to omit several unnecessary click processes as in the prior art.

In addition, the web browser receives the document for the location information list arranged in descending order from the search engine and displays it on the frame 153 of the browser screen (step 44 of FIG. 5).

Therefore, the user may be provided with information on other web pages as well as information on the web page displayed in the frame 154 using the location information list of the frame 153. At this time, the location information list displayed in the frame 153 is arranged from top to bottom in the order of high weight. Therefore, the user can grasp the weight associated with the information he wants at a glance, thereby reducing the number of unnecessary clicks.

On the frame 155 of the web browser, a search box 155a for a user to input a search word and an advertisement display unit 155b for displaying a banner advertisement are located.

Example 2

The search engine of Embodiment 1 described above uses a hyperlink to hyperlink a plurality of web pages (hyperlinks pointing to the specific web page) that are externally connected to a specific web page (web page having a default URL as an address). Extracting an index word from text and weighting the index word from the frequency of occurrence of the index word on the plurality of hyperlinked texts.

The first embodiment is an improvement over the conventional search engine that extracts all index words from the text document of the web page to be indexed.

In general, when comparing a web page with many externally linked hyperlinks and a web page with a relatively small number of externally linked hyperlinks, the incidence frequency (ie, weight) for a particular index word is relatively high. It is more likely to appear high on the web pages you have.

However, the large number of externally-linked hyperlinks tends to reflect that the web page has been out of service on the Internet for a long time, but this may represent specific information (ie, information that can represent a particular index). It does not guarantee that the document exists.

When the number of external hyperlinks is small, it is likely that the webpage has recently been serviced, but rather contains the latest information with high topicality.

Therefore, extracting and weighting index words from hyperlinks as in Embodiment 1 may not guarantee accurate search for web pages to be indexed. That is, although the phrase contained in the link text of the hyperlink is likely to represent the web page pointed to by the URL of the hyperlink, the fact that there are many hyperlinks pointing to the web page externally indicates that the web page is specific. There is no way that information has priority over other web pages. Also, it is difficult to extract all representative index words for the web page from the link text.

Example 2 was devised to compensate for this drawback of Example 1.

That is, the second embodiment supplements the disadvantages that the first embodiment may have while fully accommodating the advantages of the first embodiment.

In the second embodiment, a detailed indexing process, that is, an index word that can represent the web page is primarily selected from hyperlinks indicating a specific web page, and a key index word is selected second in the text of the web page. The index word is expanded by including the index word in the first selected index word, and the occurrence frequency of the extended index words is measured on the web page and subpages located under the tree structure of the web page to weight the index words. The process of giving.

Therefore, the second embodiment compensates for the disadvantage of the conventional disadvantage that the stopword is selected by selecting the index word from hyperlink texts indicating the web page to be indexed and the key index words on the text of the webpage. In the weighting for each index word, since the text of the web page and its subpages is based on text, reliability and fairness in weighting can be improved.

Hereinafter, the configuration of another embodiment of the search engine of the present invention will be described in detail with reference to FIGS. 7 to 11.

The technical concept shown in FIG. 2 is equally applied to this embodiment as in the first embodiment. Therefore, hereinafter, description of the technical concept shown in FIG. 2 will be omitted by referring to the description of the first embodiment.

7 is a view for explaining the configuration of the search engine according to the present embodiment.

The search engine 1000 of the present embodiment is connected to the user interface 1200 and the distributed database 1100 through a network such as the Internet.

The search engine 1000 of the present embodiment includes a hyperlink indexing unit 1500, an index expansion unit 1400, an index word weighting unit 1300, a hyperlink file 1600, a conversion file 1700, and an index file ( 1800, the query processing unit 1900.

The hyperlink index extractor 1500 extracts hyperlinks indicating a specific web page (web page having a default URL as an address) from a distributed database 1100 on the Internet, and indexes the link words from the link texts of the hyperlinks. These indexes are selected to form an index set corresponding to the corresponding default URL and then stored in the hyperlink file 1600.

The hyperlink file 1600 is a storage means for storing &quot; default URL: index word set &quot; formed in the hyperlink index index extractor.

The index word expansion unit 1100 reads one "default URL: index set" from the hyperlink file 1600, retrieves a web page having the default URL as an address from the distributed database 1100, and retrieves the web page from the corresponding web page. After selecting a key index, the key index is added to the index set to expand the index set.

The index word weighting unit 1300 receives a "default URL: extended index word set" from the index word expansion unit 1400 and positions of web pages having the default URL as an address and subpages connected to a depth N in a tree structure. After extracting the information URL based on the web page, the corresponding subpages are retrieved from the distributed database 1100 based on the location information.

In addition, the index word weighting unit 1300 extracts index words belonging to the extended index word set from the web page and the subpages of the web page having the default URL as an address, and checks the occurrence frequency of each index word. By weighting each index word according to the frequency, a "default URL: index word (weighted) set" is generated and stored in the index file 1800.

The "default URL: index word list" of the index file 1800 is converted into "index word: location information (weight value) set" and stored in the conversion file 1700.

The query processing unit 1900 receives the user query 1220 through the user interface 1200, analyzes the query, extracts an index, and then positions information corresponding to the indexes extracted from the converted file 1700. The set is read and processed according to a general search result reprocessing method, and the location information is sorted in descending order according to the weight.

In addition, the query processing unit 1900 provides the descending ordered location information list and the highest ranking location information to the user computer through the user interface 1200 as a search result.

Hereinafter, a detailed configuration of the search engine of the present invention will be described with reference to FIG. 8.

<Building Index Database>

As shown in FIG. 7, the search engine 1000 of the present embodiment includes a hyperlink index search unit 1500, an index word expansion unit 1400, an index word weighting unit 1300, a hyperlink file 1600, and a transformation. The file 1700, the index file 1800, and query processing unit 1900.

The hyperlink index extractor 1500 includes a user interface 40, a first URL storage file 1510, a storage file reset and URL allocation module 1530, a second URL storage file 1540, and a first web page. The acquisition module 1550, the hyperlink analysis module 1560, the index word extraction module 1580, and the first index word expansion module 1590.

The index word expansion unit 1400 includes a link URL obtaining module 1410, a second web page obtaining module 1420, a key index extracting module 1430, and a second index expanding module 1450. The weighting unit 1300 includes a subpage URL extraction module 1440, a subpage obtaining module 1460, and a weighting module 1310.

The query processing unit 1900 includes a query module 1920 and a query processing and ranking module 1910.

The user interface 40 is a means for the administrator of the search engine server to input an initial URL to the search engine server for building an index database. Here, the initial URL is preferably the root URL 1510. Here, the root URL refers to a URL in which a ancestor page corresponding to a hierarchical upper level of the corresponding URL does not exist in the hierarchical tree structure on the web. The tree structure also includes a directory structure containing a web page indicated by the root URL and a graph structure between web pages existing in the subdirectory and reachable along hyperlinks in the web page indicated by the root URL.

At this time, the administrator of the search engine server inputs at least one root URL 1510 through the user interface 40.

The first URL storage file 1520 is a storage means for storing the initial root URL 1510 input from the user interface 40 and the root URL 1571 transmitted from the hyperlink analysis module 1570 described below. to be.

The storage file reset and URL assignment module 1530 resets the second URL storage file 1540, reads one root URL from the first URL storage file 1520, and then stores the one URL in the second URL storage file 1540. Program module that executes commands to be stored in.

The second URL storage file 1540 stores the root URL assigned by the storage file reset and URL assignment module 1530 and stores the subpage URL 1572 transmitted by the hyperlink analysis module 1570 to be described later. It is a storage means for storing.

The first web page acquisition module 1550 is a means for interfacing the search engine 140 and an external website 120 (web server) through a network 130 such as the Internet.

That is, the first web page acquisition module 1550 reads one URL 1541 from the second URL storage file 1540, and reads the web page 1542 indicated by the read URL 1541 from the Internet. The program module is obtained from the web server 120 via the same network 130.

The first web page acquisition module 1550 transfers the web page 200 acquired through the Internet 130 to a hyperlink extraction module 1560 which will be described later.

The hyperlink extracting module 1560 is a program module that sequentially extracts all the hyperlinks existing on the obtained webpage 1542 and sequentially delivers them to the hyperlink analyzing module 1570 described later. Here, "extract and transfer the hyperlink" means determining whether a hyperlink exists on the web page 200, and obtaining a web page when the hyperlink does not exist on the web page 200. If the module 1550 is notified, and hyperlinks exist on the web page 200, the hyperlinks are extracted one by one so as not to overlap each other and the "location information and hyperlink text" 1561 of the corresponding hyperlink will be described later. Indicated to each of the analysis module 1570.

The hyperlink analysis module 1570 is a program module for analyzing the location information of the hyperlink transmitted from the hyperlink extraction module 1560. That is, by analyzing the URL of the hyperlink transmitted from the hyperlink extraction module 1560, it is not only determined whether the URL includes a root URL portion read from the first URL storage file, but also the default URL ( default) URL is also determined.

Here, the root URL and the default URL have the same meaning as in the first embodiment.

The hyperlink analysis module 1570 stores the root URL 1571 in the first URL storage file 1520 and the subpage URL 1572 through the routines of steps S109 through S124 of FIGS. 11A through 11B. 2 is stored in the URL storage file (1540).

In addition, the hyperlink analysis module 1570 determines whether the URL indicated by the extracted hyperlink is the default URL, and selects only the hyperlink having the default URL to describe the corresponding location information and link text 1573. Transfer to the index word extraction module 1580.

The index word extracting module 1580 selects an index word from the link text of the hyperlink having a default URL, and uses the selected index words so that &quot; location information: index word set &quot; 1581 [URL: Index = {a, b, c}].

The index word set formed by the index word extraction module 1580 is expanded in the first index word expansion module 1590.

That is, the first index word expansion module 1590 receives "location information: index word set" 1581 from the index word extraction module 1580 and has "location information: index word set" having the same position information as the received position information. It is determined whether or not the hyperlink file 1600 exists.

In this case, when the same location information does not exist, the received "location information: index word set" 1581 is stored in the hyperlink file 1600 as it is, and when the same location information exists, an index word corresponding to the location information. Read them.

The index words read from the hyperlink file 1600 are added to the index word set delivered by the index word extraction module, and the "location information: extended index word set" 1592 expanded by the addition of the index words is the hyperlink file 1600. The "location information: index word set" is continuously updated by being stored in.

The link URL obtaining module 1410 obtains one URL (ie, a default URL) from the hyperlink file 1600 and delivers the URL (0) 1411 to the second webpage obtaining module 1420. Module.

The second web page obtaining module 1420 receives one URL [URL (0)] 1411 from the link URL obtaining module 1410 and displays the web page 1422 indicated by the URL 1411 on the Internet. The program module obtains from the website 120 on 130.

The key index word extraction module 1430 is a program module for selecting a key index word from the text of the web page acquired by the second web page acquisition module 1420. Here, the "key index" is a word selected from a title tag or a tagged phrase used for emphasis or a word placed after a topical clue phrase or a word unregistered in a vocabulary electronic dictionary. Is defined as

The "title tag" is a tag used to indicate the title of an HTML document and appears in the title bar of the web browser. This title tag is followed by a title that indicates what this web page shows. Therefore, there is a high possibility that words (core words) capable of compressively expressing the characteristics of the web pages to be indexed in the phrases tagged with the title tag.

The term " tag used for emphasis " means a tag having an attribute of increasing the size of a letter relatively large or a tag used for underlining for emphasis. Therefore, the tagged phrase used for emphasis is likely to have a word (core word) that can represent the web page.

The clue phrase refers to a phrase that expresses a name, name, trade name, etc., for example, "name: ...", "company name: ...". Therefore, after such clue phrases, words (key words) directly related to the contents of the website are placed.

In addition, the unregistered words in the vocabulary electronic dictionary are low occurrence words but relatively high in importance.

As such, words representing the web pages to be indexed or words that are directly related to the contents of the web pages are classified as core indexes and selected from the corresponding web pages, thereby selecting index words that can appear when selecting index words only in hyperlinks. It can compensate for intolerance.

The key index words selected by the key index word extraction module 1430 are transferred to the second index word expansion module 1450.

The second index word expansion module 1450 imports an index word set corresponding to the URL (0) of the web page where the key index word is selected from the hyperlink file 1600, and adds the selected key index word to the index word set. A program module that extends the index set.

The subpage URL extraction module 1440 extracts URLs corresponding to subpages up to a depth N connected to the webpage in a tree structure based on the webpage transmitted from the second webpage obtaining module 1420. Is a program module.

9 shows a state diagram in which a web page to be indexed and subpages of the webpage are connected in a tree structure.

As shown in Fig. 9, a first layer subpage indicated by URL (1,1) and URL (1,2) exists in the first lower layer of the web page (URL (0)) to be indexed. In this way, the web pages indicated by the URL (0) have subpages of the first, second, ..., Nth hierarchies connected in a hierarchical tree structure.

The subpage URL extracting module 1440 extracts the URL (0) of the web page to be indexed and the URLs of the subpages (subpages up to depth N) of the webpage and extracts the URLs of the subpage obtaining module 1460. To pass.

The subpage obtaining module 1460 is a subpage of FIG. 9 connected to the URL [URL (0)] of the webpage to be indexed from the subpage URL extraction module 1440 and the webpage to a depth N in a tree structure. URLs (URL (1), URL (2), URL (3) ...)] of all the webpages 1462 corresponding to these URLs are obtained from the website 120 on the Internet 130. Program module.

Web pages 1463 obtained from the subpage obtaining module 1460 are transferred to the weighting module 1310.

10 illustrates a state in which the weighting module 1310 extracts an index word using the obtained web pages and weights the index word.

First, the weighting module 1310 receives the extended index word set 1451 from the index word expansion module 1451, and extracts the index word list 1312 as shown in FIG. 10 from the index word set 1451.

Using the extracted index index 1312, the weighting module 1310 is configured to provide web pages 1463 (web pages to be indexed and tree structures with the web pages to be indexed) transmitted from the subpage obtaining module 1450. Subpages connected up to depth N) and check the frequency of each index word. Based on the occurrence frequency checked in this way, a weight for each index word is calculated and &quot; location information: index word (weighted) set &quot; [URL (0): {a (w1), b (w2), c (w3)] }] 1311 is generated and stored in the index file 1800.

In addition, the weighting module 1310 may assign a higher weight to the index word selected from the link text of the hyperlink or the index word searched from the tree root webpage of the webpage tree structure than the index word searched from the subpage. .

When all of the &quot; positional information: index word (weighted) set &quot; are stored in the index file 1800, a converted file is formed by the converted file generating module 1710 described later.

The conversion file generation module 1710 stores "location information: index word (weighted) set" stored in the index file 1800 as "index word: location information (weighted) set" 1711 [a: {URL (w1), URL ( w2), and the program module converts the URL (w3).

The "index word: location information (weighted) set" 1711 [a: {URL (w1), URL (w2), URL (w3) ...]] is stored in the conversion file 1700, whereby the final index database is stored. Is generated.

An index database construction process based on the above-described configuration will be described in detail with reference to the flowcharts of FIGS. 11A to 11D.

When the search engine server manager of the present invention inputs at least one webpage root URL 1510 through the user interface 40 of the server computer 140, the root URL 1510 is the first URL storage file 1520. Are stored in. (Step S101)

The storage file reset and URL allocation module 1530 reads one root URL from the first URL storage file 1520. (Step S102) At this time, if it is impossible to read the root URL from the first URL storage file 1520, the flow advances to step S131, and if the reading is possible, the flow goes to step S104. (Step S103) Here, the inability to read the root URL from the first URL storage file means that there is no more information of the root URL to be processed.

If the root URL can be read in step S103, the storage file reset and URL assignment module 1530 resets the second URL storage file 1540 and stores the read root URL in the second URL storage file 1540. Save it. (Step S104)

The web page acquisition module 1550 reads one URL 1541 from the second URL storage file 1540. (Step S105) At this time, if it is impossible to read the URL 1541 from the second URL storage file 1540, the process returns to step S102 (step 107), and if the reading is possible, the process goes to step S108. (Step S106) Here, the inability to read the URL from the second URL storage file means that the extraction of the hyperlink information and the index word selection for all the URLs present in the second URL storage file have been completed, and therefore, the first 1 Means that you need to assign a new root URL from the URL archive and reset the existing URL.

If the URL can be read in step S106, the web page obtaining module 1550 obtains the web page 200 indicated by the URL from the web server 120 on the Internet 130 (step S108).

When the acquisition of the web page is completed, the web page is delivered to the hyperlink extracting module 1560, and the hyperlink extracting module 1560 sequentially extracts the hyperlink from the web page (step S109).

At this time, if there is no hyperlink that can be extracted on the web page (the extraction of the hyperlink is impossible), the process returns to step S105 (step S111), and if the hyperlink exists (the extraction of the hyperlink is possible). Case), the process proceeds to step S112. (Step S110) In this case, when there is no hyperlink to be extracted on the web page, when there is no hyperlink to be extracted from the web page at all and the extraction of all hyperlinks existing on the web page is completed. Includes cases.

If it is possible to extract the hyperlink from the web page in step S110, the hyperlink extraction module 1560 obtains the link URL and the link text from the extracted hyperlink information. (Step S112)

The root URL is extracted from the link URL obtained in step S112 (step 113), and it is determined whether this root URL matches the root URL in step S102. (Step S114)

At this time, if the root URLs match with each other (step S115), it is determined whether the link URL obtained in step S112 exists in the second URL storage file. (Step S116) At this time, if the link URL exists in the second storage file, the process proceeds to step S124 to determine whether the link URL is the default URL, and if not, proceeds to step S119 and transfers the link URL to the second URL storage file. Save. (Step S119)

If the root URLs do not coincide with each other in step S115, it is determined whether the root URL extracted in step S113 exists in the first URL storage file. (Step S120)

In this case, if the root URL exists, the process proceeds to step S124 to determine whether the obtained link URL is the default URL. If not, the process proceeds to step S123 and the root URL extracted in the step S113 is stored in the first URL storage file. Proceed to S124.

Steps S113 to 123 are routines for explaining a process of extending URL information of the first and second URL storage files by extracting another root URL and a subpage link URL from one root URL. At this time, it is determined whether the extracted root URL and the subpage link URL exist in the first and second storage files in order to prevent storage of duplicate URLs.

In step S124, the hyperlink analysis module determines whether the link URL obtained in step S112 is a default URL. At this time, if the link URL is not the default URL, the process returns to step S109 (step S126), and another link URL is extracted from the web page, and if it is the default URL, the process goes to step S127.

Here, the reason for selecting only the default URL among the extracted link URLs is that the web page indicated by the default URL tends to represent the sub-web pages and thus contains more important information than the sub-web pages. This is because it is possible to prevent duplication of similar search results and to move freely from the web page indicated by the default URL to the sub-web page.

In step S127, the index word extraction module selects all possible index words from the link text of the corresponding default URL and generates "location information: index set" [URL (0): {a, b, c}].

When generation of the "location information: index set" is completed in step S127, the index word extraction module generates the position of "location information: index set" [URL (0): {a, b, c}] generated in step S127. It is determined whether or not "location information: index set" having the same location information as the information exists in the hyperlink file. (Step S128)

If "Location Information: Index Set" having the same location information does not exist in the hyperlink file, the "Location Information: Index Set" generated in S127 is stored in the hyperlink file as it is and returns to Step S109. (Step S128b) If present, index words corresponding to the same positional information are read from the hyperlink file. (Step S128c)

The extended "position" in addition to the "location information: index set" [URL (0): {a, b, c}] generated in the step S127 by the index words c, d and e read in the step S128c. Information: Index Set "(URL (0): {a, b, c, d, e}] is created and stored in the hyperlink file, and the existing" Location Information: Index Set "data is rewritten. (Step S129), the flow returns to step S109. (Step S130)

If the root URL to be read from the first URL storage file no longer exists in step S102 while repeating the routines of steps S102 to S130 described above, the process proceeds to step S131 to execute the index file and the conversion file generation routine. do.

The link URL obtaining module 1410 reads one URL from the hyperlink file and delivers this URL (0) 1411 to the second webpage obtaining module 1420 (step S131).

The second web page obtaining module 1420 retrieves the web page 1422 indicated by the read URL (0) from the corresponding website 120 through the Internet 130, and loads the web page 1422 into the core. It is transmitted to the index word extraction module 1430 and the subpage URL extraction module 1440, respectively (step S132).

The core index word extracting module 1430 selects the core index words a, f, and g from the text of the obtained web page, and selects "Location information: Core index list" [URL (0): (a, f, g)]. 1143 and generate it, and transmit it to the second index extension module 1450 (step S132a).

The second index word expansion module transmits the location information [URL () of "location information: key index word list" [URL (0): ((a, f, g)]] 1431 transmitted from the core index word extraction module in the hyperlink file. "Location information: Index word set" [URL (0): {a, b, c, d, e ...]] 1610 corresponding to 0) is read.

The second index word expansion module 1450 reads the key index word list ("location information: index word set" "[URL (0): {a, b, c, d, e ...]] 1610 in this manner. a, f, g) is added to generate extended "location information: index word set" [URL (0): {a, b, c, d, e, f, g ...]] 1451 This is transmitted to the weighting module 1310. (Step S132c)

On the other hand, the subpage URL extraction module 1440 is a URL (URL (1), URL (2) of subpages from the webpage obtained in step S132 to a depth N connected to the webpage in a tree structure as shown in FIG. ), URL (3) ..], and then URL (0) of the webpage and URLs of subpages of this webpage (URL (1), URL (2), URL (3) ...)] It is transmitted to the subpage acquisition module 1460. (Step S132b)

The subpage obtaining module 1460 receives the location information [ULR (0), URL (1), URL (2), URL (3) ...] 1441 transmitted from the subpage URL extracting module 1440. The web page and the subpages 1463 of the web page are obtained from the Internet and transferred to the weighting module 1310. (Step S132d)

As shown in FIG. 10, the weighting module 1310 extends the extended "location information: index word set" transmitted from the second index word expansion module 1450 [URL (0): {a, b, c, d , e, f, g ...}] 1145 are extracted from the extended index list (a, b, c, d, e, f, g) 1312. In addition, the index words are used to check the frequency of occurrence 1464 of each index word in the texts 1463 of the web page and the subpages of the web page transmitted from the subpage obtaining module 1460, Based on "Location information: index word (weighted) set" "[URL (0): {a (w1), b (w2), c (w3) ...}] 1313 (step S135) This is stored in the index file 1800. (Step S136)

The conversion file generation module 1710 stores &quot; location information: index word (weighted) set &quot; &quot; stored in the index file [URL (0): {a (w1), b (w2), c (w3) ...] ] 1311 to "index: location information (weighted) set" "[Index = a: {URL1 (w1), URL2 (w2), URL3 (w3) ...]] (1711) The index database is formed by storing in the conversion file 1700. (Step S137)

<Search the index database>

Since the search for the index database (or the conversion file) is the same as in the above-described first embodiment, the description thereof is omitted by referring to the first embodiment.

<Display of search results>

The method of viewing the searched information on the browser screen of the user's computer using the search engine of the second embodiment is also substantially the same as that of the first embodiment (see FIG. 6) described above. Omit.

<How to save and use search results>

In general, among the information retrieved by the search engine is information that the user wants to use regularly (or periodically). Therefore, there is a feature called "Favorites" in web browsers such as Navigator and Explorer. In other words, this bookmark function stores the URL information of web pages frequently visited by users in the memory of the user's computer, and reads the URL information of web pages frequently visited by the user from the memory while calling the browser. Simply connect to a web page that supports.

However, the bookmark function of the web browser has a space limitation. As the space for accessing the Internet is diversified to users' homes, workplaces, and PC rooms, the bookmark function, which is dependent on a web browser on an individual computer, is less useful.

Therefore, the present inventor stores a user's searched information (the address of a favorite web page) as shown in FIG. 12 in a specific server on the Internet, and enjoys surfing the web without the help of a web browser even when connected to the server. The method of using search information was created.

As shown in Fig. 12, a method of using search information according to the present invention is realized through a network such as the Internet.

The search engine server 2400 having multiple processors (not necessarily multiple processors) 2430 and memory 2420 is connected to the index database 2410.

The user is provided with search information (address list of a web page corresponding to the query) by inputting a query word while accessing the search engine server 2400 through the client computer 2200. Alternatively, the user may be provided with the location information list arranged in descending order with the highest ranking location information by using the same search engine as in the first embodiment or the second embodiment from the search engine.

In this case, the user selects at least one or more location information from the location information list of the web pages provided as a search result, and clicks the MyLink button 2210 of the client computer 2200 to display the selected location information list in the MyLink server 2500. To transmit.

The location information list transmitted from the client computer 2200 is stored in the memory 2510 of the mylink server 2500. At this time, when the user is specifically linked by the MyLink server 2500 and membership registration, etc., there is a user storage area for storing the location information list of the user in the memory 2510.

Here, the mylink server includes a web server that manages the user's personal information or a web server where the user's homepage resides. Of course, the mylink server of the present invention is not necessarily limited to such a web server.

In addition, as shown in FIG. 6, when the search results are divided into frames and displayed on the browser screen of the user's computer, URLs of web pages corresponding to the highest position information are selectively (or automatically) transmitted to the MyLink server for storage. It may be.

As such, if the location information obtained by the user is stored in the MyLink server, the user may access the MyLink server and enjoy the bookmark function without the help of a browser while enjoying the services provided by the server.

The present invention is not limited to the above-described embodiments, and various modifications and variations are made by those skilled in the art within the equivalent scope of the technical concept of the present invention and the claims to be described below. Of course it is possible.

The search engine of the present invention enables quick and accurate access to information corresponding to a query while minimizing indexing.

In addition, the search engine of the present invention can non-stop displaying the web page document most relevant to the query on the browser screen to simplify the process of accessing the desired information.

In addition, the present invention stores the search information provided from the search engine in a specific web server, so that not only your own computer but also any computer that can access the Internet.

Claims (18)

A communication interface for acquiring the web page through the Internet from the location information (URL) of the worldwide web page; Hyperlink extracting means for extracting one by one all hyperlinks on the webpage acquired by the communication interface one by one; Hyperlink selecting means for selecting only hyperlinks having a default URL that can access a web page even if a file name is omitted or omitted in a uniform resource locator (URL) structure; Collecting means for collecting hyperlinks having the same default URL from a World Wide Web database; Index word extracting means for extracting an index word from the link text of the hyperlinks collected by said collecting means; As a result, a worldwide web page indexing system for indexing the web page indicated by the default URL with the extracted index terms. The method of claim 1, wherein the index word extraction means An index word extraction module for forming " location information: index list " from hyperlink information indicating a web page corresponding to the default URL; A first storage file for storing and storing the " location information: index list "; "Location information: index list (weighted list)" by searching all index list corresponding to the specific location information stored in the first storage file, counting the occurrence frequency for each index word, and weighting the index word from the occurrence frequency. An index module for indexing a web page indicated by location information; A second storage file for storing the " location information: index list "; A conversion module for converting a "location information: index list (weighted list)" of the second storage file into an "index index: location list (weighted list)"; And a third storage file for storing the "index: location information (weight) list". The method of claim 2, In assigning weights to index terms, we calculate not only the frequency of occurrence but also the length and length of the service periods of the web pages corresponding to the index terms, so that web pages with shorter service periods have higher weights than web pages with longer service periods. Indexing system for worldwide web pages, characterized in that the grant. A system that indexes web pages using a distributed database on the Internet. A hyperlink index index is obtained which obtains a web page from the distributed database based on the location information (URL), selects an index word from the link text of the hyperlink existing on the web page, and forms "location information: index set". Way; A hyperlink file for storing the "location information: index set"; Reads " location information: index set " stored in the hyperlink file, selects a key index word from text on a web page corresponding to the read location information, and indexes the index set with location information which is a URL in a hyperlink. Index word expansion means for expanding; Acquires a subpage up to a depth N connected to the webpage in a tree structure, and searches the obtained subpage and the webpage to check the frequency of occurrence of each index word of the extended index word set according to the frequency of occurrence. Index word weighting means for weighting index words of the extended index word set; And an index file for storing " location information: index set (weighted) set " weighted by the index word weighting means. The method according to claim 4, wherein the hyperlink indexing means is extracted. A web page obtaining module for obtaining a corresponding web page from the location information (URL) of the worldwide web page through the Internet; A hyperlink extracting module for sequentially extracting all the hyperlinks on the web page acquired by the web page obtaining module one by one; A hyperlink sorting module that selects only hyperlinks having a default URL that can access a corresponding web page even if a file name is omitted or omitted from a uniform resource locator (URL) structure; And an index word extracting module that selects an index word from link texts of the hyperlinks selected by the sorting module to form "location information: index word set". The method of claim 4, wherein the index expansion means A web page acquiring module for acquiring one link URL from the hyperlink storage file and importing a web page corresponding to the link URL from the distributed database; A core index word extracting module for selecting a core index word from the document text on the web page; And an index word expansion module that reads " location information: index set " having the same location information as the link URL stored in the hyperlink index file, and expands the index set by adding the core index word. Indexing system for worldwide web pages. The method of claim 6, The key index word is a word extracted from a title tag of an HTML document on a web page, a tagged phrase used for emphasis, and a phrase appearing after a clue word that is a suggestive word. Or new words or proper nouns, such as words that are not registered in the vocabulary electronic dictionary. The method of claim 7, wherein the index word weighting means A subpage URL extraction module for extracting location information of a subpage up to a depth N connected in a tree structure to a webpage acquired by the webpage obtaining module based on the obtained webpage; A subpage obtaining module for obtaining a corresponding subpage from the distributed database based on the location information of the subpage; Search all subpages and corresponding webpages acquired for the index words of "Location Information: Index Set" generated by the Index Expansion Module, and count the frequency of occurrence of the index words in the document text of the pages. Indexing system of the worldwide web page, characterized in that it comprises a weighting module for adding a weight according to the frequency of occurrence. The method of claim 8, In the weighting module assigning a weight to each index word based on the occurrence frequency of the index word, the weighting module assigns a higher weight to the index word selected from the link text of the hyperlink, and the tree root tree page of the web page tree structure. The indexing system of the worldwide web page is characterized by giving a higher weight to the index word searched in the sub-page than the index word searched in the subpage. In indexing web pages using distributed databases on the Internet, Forming " location information: index set " so that an index word for a specific web page and location information of the web page correspond to each other; Retrieving location information of the subpages up to a depth N connected in a tree structure with the location information of the "location information: index set"; Obtaining a subpage corresponding to the location information extracted in the drawing step from the distributed database; And counting the frequency of occurrence of the index words of the " location information: index set " in the web page and the sub-page of the web page, and assigning a weight to each index word according to the frequency of occurrence. How to weight it. The method of claim 10, In assigning a weight to each index word, the weighted index word is characterized by reflecting the weight of the index word as well as the importance of the structure-based search. Extracting an index word by analyzing a query word input through a user computer; Retrieving a plurality of web page location information corresponding to the index terms from the index database using the extracted index terms; Processing the retrieved location information according to the query word; Arranging the processed plurality of position information in descending order according to the weight thereof; Displaying the web page corresponding to the highest position information on the user's computer non-stop by transmitting the highest position information among the plurality of position information arranged in descending order to a web browser of the user's computer. Viewing method. The method of claim 12, And displaying a list of the plurality of location information arranged in descending order on a user computer. The method of claim 13, And a web page corresponding to the highest position information and a list of the position information are simultaneously displayed on one screen. The method of claim 14, The display screen of the user's computer is divided into frames, the web page and the location information list are simultaneously displayed in different frames, and the remaining frames are provided with a search box for inputting a query and an advertisement display unit. Viewing method. A search engine server analyzing the query words transmitted from the user computer to extract at least one or more index words, and retrieving a plurality of document location information (URLs) related to the extracted index words from the index database; Processing the retrieved location information according to the query word; Transmitting, by the search engine server, the processed location information list to a user computer; Displaying the location information list on a screen of a user computer; Selecting at least one or more desired location information from the location information list displayed on the user computer; Transmitting the selected at least one location information to a specific server linked with a user computer; And storing the location information transmitted by the specific server in a storage area of the corresponding user. Transmitting, by the search engine server, location information of the search document to a user computer; Displaying a search document on a screen of the user computer by a web browser of the user computer based on the location information; Transmitting location information of the displayed search document to a specific server linked with the user computer according to a user's decision; And storing the location information transmitted by a specific server in a storage area of the corresponding user. A plurality of user computers and a server computer are connected by using the Internet, and the server is provided with a plurality of storage areas for the plurality of user computers linked with the server, and the plurality of web documents transmitted from the user computer in the storage area. In the networking environment where the location information list is stored for each user, Connecting a user to the server computer using his computer; Retrieving a plurality of location information lists from a storage area for the user among the storage areas of the server computer and displaying them on the user computer; Accessing a web page corresponding to the location information by selecting desired location information from the displayed location information list; As a result, a user can easily surf the web simply by connecting to the server computer.